What 529 accounts of magic show

We collected 529 written accounts of magical workings, from ceremonial diaries and grimoire evocations to hoodoo interviews, séance records and forum posts, and graded how each one turned out. Of the 459 that could be graded, 243 succeeded. That is 53%, and it is the yardstick for everything on this page.

It is not the success rate of magic. People write up their wins. What the accounts can show is which things, when they appear in an account, go with success or failure more often than that yardstick.

The short version
  1. Failure leaves clearer marks than success.
  2. One fault outweighs any number of strengths.
  3. Wanting it harder buys nothing.
  4. If you repeat a working, decide the schedule beforehand.
  5. What goes with success is weaker and narrower.
  6. Most outcomes stay unexplained.

1. Failure leaves clearer marks than success

47% of graded accounts failed. When an account shows one of these, the failure rate is well above that.

When the account showsAccountsFailedShare of all failures
Stayed in an ordinary state of mindNo shift of state was worked for or reached.14100%6%
Kept gripping the resultAnxious checking, worrying, or re-doing the working out of doubt.1486%6%
Stopped before it was finishedThe working was dropped or cut short before it came to completion.4276%15%
Worked while dividedSome part of the operator held back, doubted or wanted something else.3675%12%
Only askedThe wish was put as a bare request, with nothing acted out as already so.2662%7%

The two that travel best across traditions are stopping before the working is finished and working while divided. Together they are worse than either alone: 89% of the 19 accounts showing both failed. On the calls the readers were surest of, stopping early failed 100% of the time (12 accounts) and working divided 80% (10).

A partial result counts as a failure here. That matters most for the first row: none of the 14 ordinary-state accounts succeeded outright, but 3 got part of what they asked for.

2. One fault outweighs any number of strengths

Share of accounts that succeeded, by how many strengths and faults each one shows. The small number is how many accounts are in the cell.

No faultOne faultTwo or more
No strength showing45% 10130% 237 only
One strength60% 1337 only21% 14
Two strengths57% 5430% 108% 12
Three or more73% 897 only2 only

Read across any row. A single fault pulls the rate to about four in ten however many strengths sit beside it. Strengths pay off only when no fault is present, and clearly only at three or more.

3. Wanting it harder buys nothing

Accounts that show some wanting and accounts that show strong wanting succeeded at the same rate, 60%. Wanting also does not rescue a flawed working. Where the operator wanted the result but stopped early, 17% succeeded (29 accounts). Where they wanted it but were divided, 19% (27).

4. If you repeat a working, decide the schedule beforehand

Two fresh readers went back to a wider stretch of the source for every account first read as letting go or as gripping, with the outcome cut out. They could see a distinction the first reading had blurred. Where both saw repetition on a schedule fixed in advance, 82% succeeded (17 accounts). Where both saw the working re-done out of doubt or impatience, 17% did (6).

Small counts, and most of the planned-repetition accounts are New Thought testimonials, where 67% succeed anyway. Treat this as the best lead on the page, not a settled result.

5. What goes with success is weaker and narrower

When the account showsAccountsSucceededShare of all winsFrom New Thought
Worked undividedNo part of the operator is seen holding back.9468%26%24 of 94
Lived it as already doneThe wish was acted out as a present fact.5767%16%36 of 57
Carried it to completionThe working was brought to its finish before being let go.13163%34%38 of 131
Wanted the resultSome or strong wanting is visible in the account.18560%46%48 of 185
Got a sign during the workingSomething answered while the working was under way.17960%44%2 of 179
Let go afterwardThe operator laid the matter down once the working was done.956%2%1 of 9
Had one clear aimThe working was for a single stated thing.36955%83%42 of 369

Nothing here beats the yardstick by as much as the faults do. The best-looking rows lean on New Thought testimonials, where 67% of accounts succeed whatever they show. A clear aim is nearly universal and tells you almost nothing.

6. Most outcomes stay unexplained

What these accounts cannot tell you:

How the accounts were gathered, read and scored is on the Method page. Every account, with its source, is on the Records page.